The invention relates to a system for reducing noise of recording mediums or transmission mediums and, more particularly, a noise reduction system suitable for reducing noise of sound signal recording/reproducing apparatuses such as tape-recorders.
In application of the noise reduction system to a transmission system treating the sound of high quality (including recording/reproducing system), so far as the S/N improvement of the system is not so good, the modulation of noise level is not noticeable. On the other hand, in the system with highly improved S/N, the noise level modulation is not negligible. Particularly, when it is applied to a transmission system with poor S/N, the noise level modulation is remarkable and it is practically problematic in some of sound sources.
This will be detailed by using a conventional noise reduction system with highly improved S/N with reference to FIGS. 1 and 2.
Referring now to FIG. 1, there is shown a conventional noise reduction system comprising an encoder 2 which is used in recording operation and serves as means for compressing the dynamic range of an analogue information signal, and a recording medium (or transmission medium) 1 such as tape-recorders connected at the input to the encoder 2, and a decoder 3 which is used in reproducing operation as means for expanding the compressed signal to the original dynamic range of the signal, being connected at the output of the recording medium 1.
A voltage controlled amplifier (VCA) 21 of the encoder 2 and a voltage controlled amplifier (VCA) 31 of the decoder 3 each serve as a kind of multiplier. When receiving an input signal e.sub.i and a DC level E, it produces an output signal e.sub.o expressed by: EQU e.sub.o =E.sup..+-.1 .times.e.sub.i ( 1)
In the equation (1), the figure of E takes minus sign in encoding operation while it takes plus sign in decoding operation.
A level sensor 22 of the encoder 2 and another level sensor 32 of the decoder 3 are used to detect signal level and each of the sensors 22, 32 produces the DC level E corresponding to the level of the input signal e.
Thus, if the signal e.sub.i1 is applied to the encoder 2, the encoder 2 produces an output signal (an input signal directed to the recording medium 1) e.sub.o1 is given: EQU e.sub.o1 =E.sub.o1.sup.-1 .times.e.sub.i1 ( 2)
For level expression of the signal in the equation (2), one can rewrite it to give: EQU E.sub.o1 =E.sub.o1.sup.-1 .times.E.sub.i1 ( 3)
Therefore, we obtain: EQU E.sub.o1 =E.sub.i1.sup.1/2 ( 4)
This shows that level change of the signal in encoding is compressed to 1/2 in the logarithmic scale.
When an output signal e.sub.i2 is applied to the recording medium 1, the output signal e.sub.o2 of the decoder 3 is given: EQU e.sub.o2 =E.sub.i2 .times.e.sub.i2 ( 5)
The level expression of the signal shown in the equation (5) is: EQU E.sub.o2 =E.sub.i2 .times.E.sub.i2 ( 6)
Hence: EQU E.sub.o2 =E.sub.i2.sup.2 ( 7)
This means that the level change of the signal in decoding is expanded double in logarithmic scale.
FIG. 2 shows an operation characteristic of the abovementioned noise reduction system. The compression operation of the encoder 2 operating in recording operation traces a line A. For example, the +20 dB input signal is recorded with compression of +10 dB, thus improving the peak margin. Further, the -60 dB input signal is compressed to -30 dB. Thus, the input signal is recorded with the dynamic range thereof compressed half as a whole.
The expanding operation of the decoder 3 operating in the reproduction is defined by a curve B. The signal of +10 dB recorded in the recording medium 1 is expanded to the original +20 dB. The signal of -30 dB recorded is reduced to -60 dB, as shown in the figure. Note here that, at this time noise is also reduced by 30 dB. The -60 dB signal inputted to the encoder 2 is improved of its S/N by 30 dB.
Accordingly, the improvement of S/N of the noise reduction system is generally expressed by -1/2.times.(input signal level (dB)).
In this noise reduction system, in the input signal of -100 dB, its S/N improvement reaches 50 dB so that little noise is reproduced. On the other hand, the 0 dB input signal has the S/N improvement of 0 dB. In this manner, noise is modulated in accordance with change of the loudness of the sound source.
Generally, a great loudness of the sound source masks noise so that the noise modulation is not problematic. But in the sound with relatively simple harmonic such as piano solo, noise is insufficiently masked. In this case, the level of noise changes with rhythm of the sound source. When the noise level changes greatly, it more stimulates auditory sensation level than when noise with a fixed level appears.
One of the known schemes to remove such a defect is shown in FIG. 3. Briefly, the scheme employs a de-emphasis circuit 33 with the characteristic indicated by a curve B in FIG. 4. The de-emphasis circuit 33 is used to reduce the gain in the region including much noise of high-frequencies to forcibly suppress the noise itself. The use of the de-emphasis circuit, however, attenuates the musical sound in the high-frequency range. For this reason, it is necessary to use a pre-emphasis circuit 23 with characteristic indicated by a curve A and related in inverse-function to that of the de-emphasis circuit. The emphasis at high-frequencies when it is encoded, reduces the peak margin against the saturation level of the recording medium, thus possibly resulting in distortion of high-frequencies. Therefore, the sensitivity of the level sensor must be heightened at high-frequencies in order to more greatly compress it. For this, weighting circuits 24 and 34 with characteristic indicated by a curve C in FIG. 4, are used.
In the case of recording medium with high noise level, low saturation level and narrow band width, such as small size cassette tapes, the application of this scheme to it is not advisable. Particularly, in the sound source with simple spectrum construction, the noise modulation is remarkable and further when the spectrum envelope is high level overall the frequency range, the signal at high-frequencies is not reproduced due to saturation in the high-frequency range so that the reproduced sound is not clear with poor spacial distribution of sound. FIG. 5 shows a recording/reproducing frequency characteristic of one cassette tape recorder. This indicates that the characteristic at high-frequencies more deteriorates as the level of signal is higher. In other words, when musical sound, even if it is not encoded, is recorded at high level, its sound quality is deteriorated due to the fact that the high-frequency range is not reproduced. When encoded signal of which the high-frequency range is emphasized is recorded, this deterioration of the sound quality is more remarkable.